[Rd] R 3.0, Rtools3.0,l Windows7 64-bit, and permission agony
ggrothendieck at gmail.com
Sun Apr 21 04:11:16 CEST 2013
On Sat, Apr 20, 2013 at 9:13 PM, Hadley Wickham <h.wickham at gmail.com> wrote:
>> As far as I can tell, the steps you are recommending take place in an
>> earlier build step. This would require the user who wants to do this to
>> rebuild Rtools in its entirety, which is more trouble than it is likely to
>> be worth. Especially when you can avoid the problem by using your own batch
>> script or perl script to reset the path on those relatively rare occasions
>> when you need to use Rtools. Since buiilding Rtools for a Windows machine
>> is something than CRAN does on a regular basis, why can't they just change
>> the names there (and not bother the UNIX users, and probably not even bother
>> the UNIX users who find themselves banished to the Windows wilderness).
>> Just call them "unixfind" and "unixsort" and everyone will be able to figure
>> it out....
> Because the point of Rtools is to make windows more like linux, rather
> than less?
> I really don't understand the complaints in this thread - I have
> regularly had classes of 10-20 windows users build a package with no
> prior experience, and simply downloading and installing Rtools. In my
> opinion Rtools uses the right defaults, and with a bit of extra
> scripting work (painful at first, but only needs to be done once, e.g.
> https://github.com/hadley/devtools/blob/master/R/rtools.r) it's
> painless to use from R.
The problem is that if you don't just use the PC for running R but use
it to run other programs too then any program and that utilizes
Windows batch scripts making use of find.exe or sort.exe likely won't
work if Rtools is on your path.
The fact that devtools, batchfiles and Rcpp have workarounds mitigates
it somewhat but the problem should be fixed so it works out-of-the-box
once and for all.
Statistics & Software Consulting
GKX Group, GKX Associates Inc.
email: ggrothendieck at gmail.com
More information about the R-devel